CN115751441A - Heat supply system heating station heat regulation method and system based on secondary side flow - Google Patents
Heat supply system heating station heat regulation method and system based on secondary side flow Download PDFInfo
- Publication number
- CN115751441A CN115751441A CN202211298175.6A CN202211298175A CN115751441A CN 115751441 A CN115751441 A CN 115751441A CN 202211298175 A CN202211298175 A CN 202211298175A CN 115751441 A CN115751441 A CN 115751441A
- Authority
- CN
- China
- Prior art keywords
- secondary side
- heat
- model
- feature
- water supply
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a heat regulating method of a heating power station of a heat supply system based on secondary side flow, which comprises the following steps: establishing a load prediction model of each heating power station; when the heat load change of each heating power station exceeds a threshold value, calculating the required secondary side water supply flow according to the heat load predicted value; establishing a plate exchange model of each heat station heat exchanger; establishing a data driving model between the adjusting parameters of the secondary side circulating pump and the auxiliary adjusting device of each thermal power station and the corresponding heat exchange quantity of the primary side water supply flow and temperature and the primary side and the secondary side based on the plate exchange model based on an LSTM model of a double attention mechanism; adjusting a secondary side circulating pump and an auxiliary adjusting device of each thermal power station according to the data driving model to meet the secondary side water supply flow demand value required by each thermal power station and change the heat of the primary side entering the secondary side; and correcting the adjusting parameters of the secondary side circulating pump and the auxiliary adjusting device of each heating power station to obtain the optimal adjusting parameters.
Description
Technical Field
The invention belongs to the technical field of intelligent heat supply regulation, and particularly relates to a heat regulation method of a heating power station of a heat supply system based on secondary side flow.
Background
The heat supply system produces high-temperature hot water in a heat source plant and drives the hot water to circularly flow in a primary side pipe network to convey heat energy to each heat station, heat exchange is carried out between the primary side and a secondary side in the heat station, the heat is transferred from the primary side to the secondary side, and the secondary side supplies heat to each heat user in the secondary side pipe network. Therefore, how to change the heat of the primary side entering the secondary side by adjusting the secondary side flow under the condition of not changing the circulating water flow and temperature of the primary network and the condition of not changing the hydraulic balance of the primary side, and under the condition of unchanging the primary side flow and temperature, the corresponding plate heat exchange amount is formed between the secondary side flow adjustment and the primary side flow and temperature, and the problem of ensuring the hydraulic balance of the whole network is urgently needed to be solved at present.
Based on the technical problems, a new method for regulating the heat of the heating station of the heating system based on the secondary side flow needs to be designed.
Disclosure of Invention
The invention aims to solve the technical problem of overcoming the defects of the prior art and provides a method for adjusting the heat of a heating station of a heating system based on secondary side flow, which can realize secondary side flow adjustment through a secondary side circulating pump and an auxiliary adjusting device under the condition that the flow and the temperature of a primary side are not changed, change the heat entering the secondary side from the primary side, namely form corresponding heat exchange amount of plate exchange with the flow and the temperature of the primary side through the adjusting action of the circulating pump and the auxiliary adjusting device, and ensure the requirement of the heat of the secondary side of the heating station.
In order to solve the technical problems, the technical scheme of the invention is as follows:
the invention provides a heat supply system heating power station heat adjusting method based on secondary side flow, which comprises the following steps:
s1, acquiring historical operating data and weather data of each heating power station, and establishing a load prediction model of each heating power station after meteorological pattern clustering, characteristic importance evaluation and model training are carried out on the acquired data through a prediction model;
s2, when the heat load change of each heating power station exceeds a threshold value, calculating the required secondary side water supply flow according to a heat load predicted value;
s3, establishing plate exchange models of heat exchangers of all heating power stations;
s4, establishing adjusting parameters of a secondary side circulating pump and an auxiliary adjusting device of each thermal power station and a corresponding data driving model comprising the water supply flow and temperature at the primary side and the heat exchange quantity between the primary side and the secondary side based on a plate exchange model based on an LSTM model of a double attention mechanism;
s5, adjusting the secondary side circulating pumps and the auxiliary adjusting devices of the thermal power stations according to the data driving model to meet the secondary side water supply flow demand required by the thermal power stations and change the heat of the primary side entering the secondary side;
and S6, analyzing the operation condition after the secondary side flow is regulated according to the established heat supply system secondary network simulation model, and then correcting the regulation parameters of the secondary side circulating pump and the auxiliary regulation device of each thermal station to obtain the optimal regulation parameters of the secondary side circulating pump and the auxiliary regulation device.
Further, the step S1 includes:
respectively acquiring historical operating data and weather data according to a heat meter arranged in a heating station and a weather data interface butted with a heating system, wherein the historical operating data at least comprises secondary side water supply flow, secondary side water supply and return temperature, secondary side water supply and return pressure and heat load, and the weather data at least comprises temperature, humidity, wind speed and illumination;
taking the historical operating data and the weather data as model samples, clustering GMMs by adopting a Gaussian mixture model, and carrying out meteorological feature clustering analysis according to the self properties of the historical weather data to obtain a plurality of meteorological modes;
and selecting the characteristics of the data in the model sample by adopting a random forest algorithm, selecting a characteristic subset with higher characteristic importance, inputting the selected characteristic subset into an optimized LSSVM model, training the data in each meteorological mode, establishing a thermal station load prediction model, and superposing corresponding load prediction results in each meteorological mode to obtain a final thermal station load prediction result.
Further, the implementation process of selecting the feature subset with higher feature importance by selecting the features of the data in the model sample by using the random forest algorithm includes:
the random forest comprises a plurality of decision trees, the feature importance is calculated according to the contribution rate of each feature in each decision tree, and the feature importance of the feature is obtained by averaging the contribution rates of one feature in all the decision trees; the contribution rate is obtained by calculating a kini coefficient, the feature importance of the jth feature in the node a is calculated according to the change of the kini index, and the feature importance is expressed as follows: VIM ja ＝GI a -GI b -GI c ；GI a 、GI b And GI c The node a and the node a are respectively used for generating a Gini coefficient of two new nodes b and c after branching;
assuming there are n trees in the random forest, the importance of the jth feature on all trees is expressed as:
is the sum of the importance of features on n trees; VIM ij Is the sum of the importance of the jth feature in the ith tree;
averaging the sum of the importance of the jth feature, which is the feature importance of the jth feature,
p =1,2,3, \ 8230for the sum of all the importance of m features on n trees;
sorting all the features from large to small in feature importance, and selecting the features with high feature importance, namely the top n ranked features as a feature subset;
when the GMM carries out meteorological mode clustering, parameter estimation of mean value and covariance of the initialized GMM is carried out by adopting an expectation maximization algorithm; the expectation maximization algorithm comprises the expectation steps of: the cluster number of the GMM model is required to be set, the pre-estimated values of the mean value and covariance of the initialized GMM are solved, and the probability that the weather data only belong to the corresponding cluster is calculated; a maximization step: dividing the data points into clusters with higher probability by using a maximum likelihood function, and updating the mean value and covariance of GMM; finally, circularly performing the expectation step and the maximization step until the parameters are converged or the likelihood function is converged to obtain a clustering result of the meteorological model;
the optimized LSSVM model adopts a metaheuristic optimization algorithm AOA to optimize a kernel parameter sigma and a regularization parameter gamma of the LSSVM model, and comprises the following steps: initializing parameters of an AOA optimization algorithm, including population number, maximum iteration number, local development precision and an acceleration function; randomly generating a population, setting initial position parameters (sigma, gamma) of the population, calculating individual fitness values according to root-mean-square errors, and comparing the fitness values to obtain the current optimal population position; judging whether the initial population enters an exploration phase or a development phase, and updating the position of the initial population; comparing the updated population, taking the lowest fitness as an optimal population position, and judging whether an iteration condition is met; and performing model prediction by taking the optimal value generated by the iteration end as the parameter (sigma, gamma) of the LSSVM.
Further, the step S2 includes: when the predicted value of the heat load of each heating power station is compared with the current value of the heat load, and when the change of the heat load value exceeds a set threshold value, calculating the required secondary side water supply flow according to the predicted value of the heat load and the set value of the secondary side water supply and return temperature; otherwise, the current system operation condition is maintained.
Further, in step S3, establishing a plate exchange model of each heat station heat exchanger includes: the method comprises the steps of training a plate exchange model by adopting a neural network algorithm, and establishing a heat exchange quantity model of the heat exchanger by fitting different secondary side water supply flows, different heat exchange quantities and different primary side water supply flows and water supply temperature data of the heat station heat exchanger to describe the relationship between the secondary side water supply flow and the heat acquired from the primary side through the heat exchanger under the condition of giving the primary side water supply flow and the water supply temperature.
Further, the step S4 includes:
through the different secondary side circulating pump of each heating power station of fitting and the corresponding primary side water supply flow and temperature of auxiliary regulating device regulating parameter, the heat that the secondary side was obtained from the primary side through the heat exchanger and other historical data that influence the relevant parameter of secondary side flow, establish secondary side flow control model for the description is at certain model input vector: the method comprises the following steps that (1) under the conditions of primary side water supply flow and temperature, heat obtained by a secondary side from a primary side through a heat exchanger and other related parameters influencing secondary side flow, a model outputs required secondary side flow adjusting parameters including adjusting parameters of a secondary side circulating pump and an auxiliary adjusting device;
the auxiliary adjusting device comprises a water mixing device arranged on the secondary side of the heating station, the flow of the water mixing pipe is adjusted by adjusting the opening degree of a valve of the water mixing device, the partial flow of the secondary side backwater is introduced into the water mixing pipe, meanwhile, the secondary side circulation flow is adjusted by adjusting the parameters of a secondary side circulating pump, and the secondary side water supply flow is adjusted by the secondary side circulating pump and the water mixing device together; the other parameters related to the secondary side flow comprise a secondary side water supply and return pressure difference, a secondary side circulating pump attribute, an auxiliary adjusting device attribute and a secondary side water supply and return pressure;
the training of the secondary side flow control model adopts an LSTM model based on a dual attention mechanism, and comprises the following steps: inputting a vector layer, a characteristic attention layer, an LSTM network, a time sequence attention layer and a full connection layer for outputting; the input vector layer obtains an input feature vector x consisting of a secondary side flow regulation historical sequence and a related input feature sequence t And transmitting the input feature vector x to the feature attention layer, and performing feature extraction through dynamic distribution of feature attention weight to obtain weighted and corrected input feature vector x t '; constructing an LSTM network layer structure, extracting hidden time sequence correlation information from the weighted and corrected input feature vector, and obtaining the hidden layer state h of each historical moment t (ii) a Mining the correlation between the information of each moment of the relevant characteristic time sequence and the current moment data through the time sequence attention layer, and calculating the influence weight l of the output information of each historical moment by adopting a time sequence attention mechanism t Finally, the global hidden layer state h of each historical time information is used t Inputting the data to a full connection layer, and outputting a secondary side flow regulation predicted value y of future n steps t+n 。
Further, the feature attention layer inputs a feature vector x t As an input of the feature attention mechanism, attention weight calculation is performed on m features at the current time, which is expressed as: e.g. of the type t ＝σ(W e x t +b e )；x t ＝[x 1,t ,x 2,t ,…,x m,t ]；e t ＝[e 1,t ,e 2,t ,…,e m,t ](ii) a σ () is a Sigmoid activation function; w e Is a trainable weight matrix; b is a mixture of e A bias vector for computing feature attention weights;
normalizing each feature attention weight coefficient to obtain a feature attention weight alpha t ＝[α 1,t ,α 2,t ,…,α m,t ]The attention weight value of the mth feature is expressed as:
feature attention gained through the current timeHeavy alpha t And input feature vector x t Computing a weighted modified input feature vector x t ', indicated as: x is the number of t ′＝α t ⊙x t ＝[α 1,t x 1,t α 2,t x 2,t … α m,t x m,t ](ii) a As an hadamard product;
the feature attention layer includes, in addition to a feature attention mechanism, a CNN network for performing local feature extraction on input data before calculating attention weights.
Further, the temporal attention layer hides the layer state h of the LSTM network at time t t As an input to the time-series attention mechanism, a time attention weight corresponding to each historical time at the current time is calculated, and is expressed as: l t ＝ReLU(W d h t +b d )；h t ＝[h 1,t ,h 2,t ,…,h k,t ](ii) a k is the input sequence time window length; l t ＝[l 1,t ,l 2,t ,…,l k,t ](ii) a ReLU () is the activation function; w d Is a trainable weight matrix; b d A bias vector for computing a temporal attention weight;
normalizing the attention weight coefficient of each time to obtain the attention weight beta of the time t ＝[β 1,t ,β 2,t ,…,β k,t ]The attention weight value at the kth time is expressed as:
temporal attention weight β obtained by the current time instant t And hidden layer state h t Computing global hidden layer states h t ', is represented as:
is a matrix product.
Further, in step S6, after analyzing the operation condition after the secondary side flow is adjusted according to the established heat supply system secondary network simulation model, correcting the adjustment parameters of the secondary side circulating pump and the auxiliary adjustment device of each thermal power station to obtain the optimal adjustment parameters of the secondary side circulating pump and the auxiliary adjustment device, including:
analyzing the system operation hydraulic working condition after flow regulation is carried out according to the regulation parameters of the secondary side circulating pump and the water mixing device according to a pre-established simulation model of the secondary side network of the heat supply system, and correcting the regulation parameters of the secondary side circulating pump and the auxiliary regulation device of each heating power station according to hydraulic power dispatching loss to obtain the optimal regulation parameters of the secondary side circulating pump and the auxiliary regulation device;
the hydraulic power misscheduling is the ratio of the actual secondary side water supply flow to the predicted secondary side water supply flow, and if the hydraulic power misscheduling is larger than 1, the actual secondary side water supply flow of the heating power station is larger than the predicted flow; if the hydraulic power failure rate is less than 1, the actual secondary side water supply flow of the heating power station is less than the predicted flow;
adjusting and correcting different floating values of adjusting parameters of each heating power station secondary side circulating pump and each auxiliary adjusting device according to hydraulic power failure scheduling to obtain multiple adjusting schemes, taking the adjusting parameters in different adjusting schemes as control quantities, outputting heating power station secondary side water supply flow corresponding to different adjusting schemes after the adjusting parameters are operated through a constructed secondary network simulation model, taking the variance of the heating power station secondary side water supply flow as a hydraulic power balance standard, and calculating by adopting a global optimization algorithm to obtain an optimal adjusting scheme.
The invention also provides a heat supply system heating power station heat regulating system based on secondary side flow, which comprises:
the heat load prediction module is used for acquiring historical operating data and weather data of each heat station, and establishing a load prediction model of each heat station after meteorological pattern clustering, characteristic importance evaluation and model training are carried out on the acquired data through the prediction model;
the secondary side flow demand module is used for calculating the required secondary side water supply flow according to the heat load predicted value when the heat load change of each heating power station exceeds a threshold value;
the plate exchange model establishing module is used for establishing a plate exchange model of each heat station heat exchanger;
the secondary side flow control module is used for establishing adjusting parameters of a secondary side circulating pump and an auxiliary adjusting device of each thermal power station and a corresponding data driving model comprising primary side water supply flow and temperature and heat exchange quantity between the primary side and the secondary side based on the plate exchange model based on an LSTM model of a double attention mechanism;
the execution module is used for adjusting the secondary side circulating pump and the auxiliary adjusting device of each heating power station according to the data driving model, meeting the secondary side water supply flow demand value required by each heating power station and changing the heat of the primary side entering the secondary side;
and the adjusting parameter correcting module is used for correcting the adjusting parameters of the secondary side circulating pump and the auxiliary adjusting device of each heating power station after analyzing the operation condition of the secondary side flow adjusted according to the established heat supply system secondary network simulation model to obtain the optimal adjusting parameters of the secondary side circulating pump and the auxiliary adjusting device.
The invention has the beneficial effects that:
(1) The method comprises the steps of acquiring historical operating data and weather data of each heating power station, performing characteristic importance evaluation and model training on the acquired data through a prediction model, and then establishing a load prediction model of each heating power station; when the heat load change of each heating power station exceeds a threshold value, calculating the required secondary side water supply flow according to the heat load predicted value; establishing a plate exchange model of each heat station heat exchanger; based on an LSTM model of a double attention mechanism, establishing a data driving model between the adjusting parameters of a secondary side circulating pump and an auxiliary adjusting device of each thermal power station and the corresponding heat exchange quantity comprising the primary side water supply flow and temperature and the primary side and the secondary side based on a plate exchange model; adjusting a secondary side circulating pump and an auxiliary adjusting device of each thermal power station according to the data driving model to meet the secondary side water supply flow demand value required by each thermal power station and change the heat of the primary side entering the secondary side; analyzing the operation condition after the secondary side flow is regulated according to the established heat supply system secondary network simulation model, and then correcting the regulation parameters of the secondary side circulating pump and the auxiliary regulation device of each heating power station to obtain the optimal regulation parameters of the secondary side circulating pump and the auxiliary regulation device; the secondary side flow regulation can be realized through the secondary side circulating pump and the auxiliary regulating device under the condition that the flow and the temperature of the primary side are not changed and the hydraulic balance of the primary side is not changed, so that the heat of the primary side entering the secondary side is changed, namely, the heat exchange amount of corresponding plate exchange is formed between the flow and the temperature of the primary side through the regulating actions of the circulating pump and the auxiliary regulating device, and the requirement of the heat of the secondary side of the thermal power station is met;
(2) The method comprises the steps of evaluating the feature importance of data in a model sample and calculating a feature importance score; clustering GMMs by adopting a Gaussian mixture model, and performing meteorological feature clustering analysis according to the self property of historical weather data to obtain a plurality of meteorological modes; the accuracy of load prediction can be improved through meteorological feature clustering and feature importance evaluation;
(3) The invention well solves the problems that the relevance of each input characteristic in actual operation is ignored and the performance is poor in processing historical information in the flow prediction process of the LSTM model by adopting the LSTM model based on the double attention mechanism to train the secondary side flow control model. The characteristic attention mechanism excavates the correlation between the flow and the influence factors thereof, optimizes the input of an LSTM model and improves the overall prediction precision; the time sequence attention mechanism excavates the relevance degree of the current time flow and the historical key time information, optimizes the output of the LSTM model and improves the prediction precision of the key time point.
Additional features and advantages will be set forth in the description which follows, and in part will be obvious from the description, or may be learned by the practice of the invention. The objectives and other advantages of the invention will be realized and attained by the structure particularly pointed out in the written description and drawings.
In order to make the aforementioned and other objects, features and advantages of the present invention comprehensible, preferred embodiments accompanied with figures are described in detail below.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, and it is obvious that the drawings in the following description are some embodiments of the present invention, and other drawings can be obtained by those skilled in the art without creative efforts.
Fig. 1 is a schematic flow chart of a method for regulating heat of a heating power station of a heating system based on secondary side flow according to the present invention;
FIG. 2 is a flow chart of the heating power station load prediction method based on meteorological patterns and feature importance of the present invention;
FIG. 3 is a schematic diagram of an LSTM model based on a double attention mechanism according to the present invention;
fig. 4 is a schematic structural diagram of a heat regulating system of a heating power station of a heating system based on secondary side flow.
Detailed Description
To make the objects, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
Example 1
Fig. 1 is a schematic flow chart of a method for regulating heat of a heating power station of a heating system based on secondary side flow according to the present invention.
As shown in fig. 1, the present embodiment 1 provides a method for adjusting heat quantity of a heating system thermal station based on secondary side flow, where the method for adjusting heat quantity of a heating system thermal station includes:
s1, acquiring historical operating data and weather data of each heating power station, and establishing a load prediction model of each heating power station after meteorological pattern clustering, characteristic importance evaluation and model training are carried out on the acquired data through a prediction model;
s2, when the heat load change of each heating power station exceeds a threshold value, calculating the required secondary side water supply flow according to a heat load predicted value;
s3, establishing a plate exchange model of each heat station heat exchanger;
s4, establishing adjusting parameters of a secondary side circulating pump and an auxiliary adjusting device of each thermal power station and a corresponding data driving model comprising the water supply flow and temperature at the primary side and the heat exchange quantity between the primary side and the secondary side based on a plate exchange model based on an LSTM model of a double attention mechanism;
s5, adjusting the secondary side circulating pumps and the auxiliary adjusting devices of the thermal power stations according to the data driving model to meet the secondary side water supply flow demand required by the thermal power stations and change the heat of the primary side entering the secondary side;
and S6, analyzing the operation condition after the secondary side flow is adjusted according to the established heat supply system secondary network simulation model, and then correcting the adjusting parameters of the secondary side circulating pump and the auxiliary adjusting device of each heating power station to obtain the optimal adjusting parameters of the secondary side circulating pump and the auxiliary adjusting device.
The secondary side water supply flow is a flow interval range, the upper limit and the lower limit of flow regulation are set, the minimum flow under hydraulic balance can be usually selected, and the heat of the primary side entering the secondary side can be changed through the secondary side flow regulation without changing the hydraulic balance state of the primary side, the flow and the temperature of the primary side water supply.
Fig. 2 is a flow chart of the method for predicting the load of the thermal station based on the meteorological model and the characteristic importance.
As shown in fig. 2, in the present embodiment, the step S1 includes:
respectively acquiring historical operation data and weather data according to a heat meter deployed in a heating station and a weather data interface butted with a heating system, wherein the historical operation data at least comprises secondary side water supply flow, secondary side water supply and return temperature, secondary side water supply and return pressure and heat load, and the weather data at least comprises temperature, humidity, wind speed and illumination;
taking the historical operating data and the weather data as model samples, clustering GMMs by adopting a Gaussian mixture model, and carrying out meteorological feature clustering analysis according to the self properties of the historical weather data to obtain a plurality of meteorological modes;
and selecting the characteristics of the data in the model sample by adopting a random forest algorithm, selecting a characteristic subset with higher characteristic importance, inputting the selected characteristic subset into an optimized LSSVM model, training the data in each meteorological mode, establishing a thermal station load prediction model, and superposing corresponding load prediction results in each meteorological mode to obtain a final thermal station load prediction result.
In this embodiment, the implementation process of selecting a feature subset with higher feature importance by performing feature selection on data in a model sample by using a random forest algorithm includes:
the random forest comprises a plurality of decision trees, the feature importance is calculated according to the contribution rate of each feature in each decision tree, and the feature importance of the feature is obtained by averaging the contribution rates of one feature in all the decision trees; the contribution rate is obtained by calculating a kini coefficient, the feature importance of the jth feature in the node a is calculated according to the change of the kini index, and the contribution rate is expressed as: VIM ja ＝GI a -GI b -GI c ；GI a 、GI b And GI c The node b and the node c are respectively the node a and the node a which are generated after branching;
assuming there are n trees in the random forest, the importance of the jth feature on all trees is expressed as:
is the sum of the importance of features on n trees; VIM ij The j is the sum of the importance of the j characteristic in the ith tree;
for the jth bitThe sum of the importance of the features is averaged, namely the feature importance of the jth feature,
p =1,2,3, \ 8230for the sum of all the importance of m features on n trees;
sorting all the features from large to small in feature importance, and selecting the features with high feature importance, namely the top n ranked features as a feature subset;
when the GMM carries out meteorological mode clustering, parameter estimation of mean value and covariance of the initialized GMM is carried out by adopting an expectation maximization algorithm; the expectation-maximization algorithm comprises the expectation steps of: the cluster number of the GMM model is required to be set, the pre-estimated values of the mean value and covariance of the initialized GMM are solved, and the probability that the weather data only belong to the corresponding cluster is calculated; a maximization step: dividing the data points into clusters with higher probability by using a maximum likelihood function, and updating the mean value and covariance of GMM; finally, circularly performing the expectation step and the maximization step until the parameters are converged or the likelihood function is converged to obtain a clustering result of the meteorological model;
after a plurality of meteorological modes are obtained, performing characteristic importance evaluation and characteristic importance value calculation on weather data and running data in each meteorological mode by adopting an XGboost learning algorithm, inputting selected characteristic data into an optimized LSSVM (least squares support vector machine) model, training the data in each meteorological mode, establishing a heating station load prediction model, and overlapping corresponding load prediction results in each meteorological mode to obtain a final heating station load prediction result;
the method for optimizing the LSSVM model comprises the following steps of optimizing a kernel parameter sigma and a regularization parameter gamma of the LSSVM model by adopting a metaheuristic optimization algorithm AOA, wherein the method comprises the following steps: initializing parameters of an AOA optimization algorithm, including population number, maximum iteration number, local development precision and an acceleration function; randomly generating a population, setting initial position parameters (sigma, gamma) of the population, calculating individual fitness values according to the root-mean-square error, and then comparing the fitness values to obtain the current optimal population position; judging whether the initial population enters an exploration phase or a development phase, and updating the position of the initial population; comparing the updated population, taking the population with the lowest fitness as an optimal population position, and judging whether an iteration condition is met; and performing model prediction by taking the optimal value generated by the iteration end as the parameter (sigma, gamma) of the LSSVM.
It should be noted that GMM clustering is to characterize the cluster class of each sample based on multidimensional GMM, the result generated after clustering is a series of probability values, the individual in the sample has corresponding probabilities to different classes, and the class to which the maximum probability belongs is selected as the classification basis. The weather data at least comprises temperature, humidity, wind speed and illumination, and the GMM divides the historical weather elements into a plurality of meteorological modes according to the self properties of the historical weather elements to realize refined grouping training and prediction. From the prediction perspective, the difference of the heat load prediction under different meteorological conditions is considered in different meteorological modes, and each meteorological mode belongs to the multivariate Gaussian distribution of different parameters. For example, meteorological model 1 represents medium wind speed, medium and high air temperature, medium and high light, medium and high humidity weather conditions; the meteorological model 2 represents weather conditions of medium and low wind speed, medium and low air temperature, medium and low humidity and medium and low illumination; the meteorological model 3 represents medium wind speed, low air temperature, medium and low humidity and low light weather conditions.
The AOA optimization algorithm comprises the following steps:
(1) Selecting an optimization strategy through a mathematical optimizer acceleration function MOA, wherein the optimization strategy comprises a global exploration phase and a local development phase when a random number r 1 When the random number r is greater than MOA, a global exploration phase is carried out 1 If the number is less than MOA, a local development stage is carried out;
M max 、M min the maximum value and the minimum value of the MOA are respectively; t is the current iteration number; t is the total iteration number;
(2) In the exploration stage, global exploration is realized through multiplication and division strategies, the global optimization capability is enhanced, the premature convergence capability is overcome, and when a random number r 2 If less than 0.5, executing division search strategy, ifRandom number r 2 When the number of the multiplication is more than or equal to 0.5, executing a multiplication exploration strategy;
xi is a minimum value; u is a control parameter for adjusting the searching process; UB and LB are upper and lower boundaries of variable; MOP is the probability of a mathematical optimizer;
(3) In the development stage, local exploration is realized through addition and subtraction strategies, the local optimization capability is enhanced, and when a random number r is used 3 When the random number r is less than 0.5, executing a subtraction exploration strategy 3 When the value is more than or equal to 0.5, executing an addition exploration strategy;
by adopting the AOA with strong optimization capability and high convergence rate to optimize the two parameters of the LSSVM model, the problem that the LSSVM parameters are difficult to determine is solved, the optimization performance is better than that of the conventional optimization algorithm, and the prediction accuracy can be greatly improved.
In this embodiment, the step S2 includes: when the predicted value of the heat load of each heating power station is compared with the current value of the heat load, and when the change of the heat load value exceeds a set threshold value, calculating the required secondary side water supply flow according to the predicted value of the heat load and the set value of the secondary side water supply and return temperature; otherwise, the current system operation condition is maintained.
In this embodiment, in step S3, establishing a plate exchange model of each heat station heat exchanger includes: the method comprises the steps of training a plate exchange model by adopting a neural network algorithm, and establishing a heat exchange quantity model of the heat exchanger by fitting different secondary side water supply flows, different heat exchange quantities and different primary side water supply flows and water supply temperature data of the heat station heat exchanger to describe the relationship between the secondary side water supply flow and the heat acquired from the primary side through the heat exchanger under the condition of giving the primary side water supply flow and the water supply temperature.
Fig. 3 is a schematic structural diagram of an LSTM model based on a dual attention mechanism according to the present invention.
As shown in fig. 3, in the present embodiment, the step S4 includes:
through the heat that the heat exchanger obtained from the primary side and the historical data of other relevant parameters that influence the secondary side flow through the heat exchanger of the corresponding primary side water supply flow and temperature, secondary side that different secondary side circulating pumps of each heating power station and auxiliary regulating device regulating parameter of fitting, establish secondary side flow control model for the description is at certain model input vector: the method comprises the following steps that under the conditions of primary side water supply flow and temperature, heat obtained by a secondary side from a primary side through a heat exchanger and other related parameters influencing secondary side flow, a model outputs required secondary side flow adjusting parameters including adjusting parameters of a secondary side circulating pump and an auxiliary adjusting device;
the auxiliary adjusting device comprises a water mixing device arranged on the secondary side of the thermal station, the flow of the water mixing pipe is adjusted by adjusting the opening degree of a valve of the water mixing device, the partial flow of the secondary side backwater is introduced into the water mixing pipe, the secondary side circulating flow is adjusted by adjusting the parameters of a secondary side circulating pump, and the secondary side water supply flow is adjusted by the secondary side circulating pump and the water mixing device together; the other parameters related to the secondary side flow comprise a secondary side water supply and return pressure difference, a secondary side circulating pump attribute, an auxiliary adjusting device attribute and a secondary side water supply and return pressure;
the training of the secondary side flow control model adopts an LSTM model based on a dual attention mechanism, and comprises the following steps: inputting a vector layer, a characteristic attention layer, an LSTM network, a time sequence attention layer and a full connection layer for outputting; the input vector layer obtains an input feature vector x consisting of a secondary side flow regulation historical sequence and a related input feature sequence t And transmitting the weighted input feature vector x to the feature attention layer, and performing feature extraction through dynamic distribution of feature attention weight to obtain a weighted and corrected input feature vector x t '; building LSTM network layerThe structure extracts hidden time sequence related information from the weighted and corrected input characteristic vector to obtain the hidden layer state h of each historical moment t (ii) a Mining the correlation between the time information of the relevant characteristic time sequence and the current time data through the time sequence attention layer, and calculating the influence weight l of the output information of each historical time by adopting a time sequence attention mechanism t Finally, the global hidden layer state h of each historical time information is used t Inputting the data to a full connection layer, and outputting a secondary side flow regulation predicted value y of n steps in the future t+n 。
In this embodiment, the feature attention layer inputs a feature vector x t As an input of the feature attention mechanism, attention weight calculation is performed on m features at the current time, which is expressed as: e.g. of a cylinder t ＝σ(W e x t +b e )；x t ＝[x 1,t ,x 2,t ,…,x m,t ]；e t ＝[e 1,t ,e 2,t ,…,e m,t ](ii) a Sigma () is a Sigmoid activation function; w is a group of e Is a trainable weight matrix; b is a mixture of e A bias vector for computing feature attention weights;
normalizing each feature attention weight coefficient to obtain a feature attention weight alpha t ＝[α 1,t ,α 2,t ,…,α m,t ]The attention weight value of the mth feature is expressed as:
feature attention weight alpha obtained by the current time instant t And input feature vector x t Computing a weighted modified input feature vector x t ', indicated as: x is the number of t ′＝α t ⊙x t ＝[α 1,t x 1,t α 2,t x 2,t … α m,t x m,t ](ii) a An h-hadamard product;
the feature attention layer includes, in addition to a feature attention mechanism, a CNN network for performing local feature extraction on input data before calculating attention weights.
In this embodiment, the temporal attention layer hides the layer state h of the LSTM network at time t t As an input to the time-series attention mechanism, a time attention weight corresponding to each historical time at the current time is calculated, and is expressed as: l. the t ＝ReLU(W d h t +b d )；h t ＝[h 1,t ,h 2,t ,…,h k,t ](ii) a k is the input sequence time window length; l t ＝[l 1,t ,l 2,t ,…,l k,t ](ii) a ReLU () is the activation function; w is a group of d Is a trainable weight matrix; b d A bias vector for computing a temporal attention weight;
normalizing the attention weight coefficient of each time to obtain the attention weight beta of the time t ＝[β 1,t ,β 2,t ,…,β k,t ]The attention weight value at the kth time is expressed as:
the temporal attention weight β obtained by the current time t And hidden layer state h t Computing a global hidden layer state h t ', indicated as:
is a matrix product.
It should be noted that the attention mechanism is a model simulating the attention of the human brain, and the attention of the human brain to things at a certain moment is focused on a certain place for reference, and the attention of the human brain to other parts is reduced or even ignored. Attention is paid to endowing different weights to the input features of the model, more key influence factors are highlighted, and the model is further helped to make more accurate judgment. In the secondary side flow prediction, the importance degree of different input parameters on prediction information is analyzed through a characteristic attention mechanism, the weight of the secondary side flow influence factor input characteristic influence force is quantized, key characteristics are highlighted, and the characteristics with small correlation degree are weakened. The method comprises the steps that the problems that secondary side flow is greatly influenced by historical states and the influence of the secondary side flow at different moments is different are fully considered through a time sequence attention mechanism at the secondary side flow prediction end, the influence degree of state information at each historical moment on a current flow prediction result is quantized, historical state information is processed in a self-adaptive mode, and the influence of state information at related moments is strengthened.
In this embodiment, in step S6, after analyzing the operation condition after the secondary side flow is adjusted according to the established heat supply system secondary network simulation model, the adjusting parameters of the secondary side circulating pump and the auxiliary adjusting device of each thermal station are corrected to obtain the optimal adjusting parameters of the secondary side circulating pump and the auxiliary adjusting device, which includes:
analyzing the system operation hydraulic working condition after flow regulation is carried out according to the regulation parameters of the secondary side circulating pump and the water mixing device according to a pre-established simulation model of the secondary side network of the heat supply system, and correcting the regulation parameters of the secondary side circulating pump and the auxiliary regulation device of each heating power station according to hydraulic power dispatching loss to obtain the optimal regulation parameters of the secondary side circulating pump and the auxiliary regulation device;
the hydraulic power misscheduling is the ratio of the actual secondary side water supply flow to the predicted secondary side water supply flow, and if the hydraulic power misscheduling is larger than 1, the actual secondary side water supply flow of the heating power station is larger than the predicted flow; if the hydraulic power loss schedule is less than 1, the actual secondary side water supply flow of the heating power station is less than the predicted flow;
adjusting and correcting different floating values of adjusting parameters of each heating power station secondary side circulating pump and each auxiliary adjusting device according to hydraulic power failure scheduling to obtain multiple adjusting schemes, taking the adjusting parameters in different adjusting schemes as control quantities, outputting heating power station secondary side water supply flow corresponding to different adjusting schemes after the adjusting parameters are operated through a constructed secondary network simulation model, taking the variance of the heating power station secondary side water supply flow as a hydraulic power balance standard, and calculating by adopting a global optimization algorithm to obtain an optimal adjusting scheme.
Example 2
Fig. 4 is a schematic structural diagram of a heat regulating system of a heating power station of a heating system based on secondary side flow according to the present invention.
As shown in fig. 4, the present embodiment 2 proposes a heat regulating system of a heating system thermal station based on a secondary side flow, where the heat regulating system of the heating system thermal station includes:
the heat load prediction module is used for acquiring historical operating data and weather data of each heat station, and establishing a load prediction model of each heat station after meteorological pattern clustering, characteristic importance evaluation and model training are carried out on the acquired data through the prediction model;
the secondary side flow demand module is used for calculating the required secondary side water supply flow according to the heat load predicted value when the heat load change of each heating power station exceeds a threshold value;
the plate exchange model establishing module is used for establishing a plate exchange model of each heat station heat exchanger;
the secondary side flow control module is used for establishing adjusting parameters of a secondary side circulating pump and an auxiliary adjusting device of each thermal power station and a corresponding data driving model comprising primary side water supply flow and temperature and heat exchange quantity between the primary side and the secondary side based on the plate exchange model based on an LSTM model of a double attention mechanism;
the execution module is used for adjusting the secondary side circulating pump and the auxiliary adjusting device of each heating power station according to the data driving model, meeting the secondary side water supply flow demand value required by each heating power station and changing the heat of the primary side entering the secondary side;
and the adjusting parameter correcting module is used for correcting the adjusting parameters of the secondary side circulating pump and the auxiliary adjusting device of each heating power station after analyzing the operation condition of the secondary side flow after being adjusted according to the established heat supply system secondary network simulation model, so as to obtain the optimal adjusting parameters of the secondary side circulating pump and the auxiliary adjusting device.
In the several embodiments provided in the present application, it should be understood that the disclosed system and method may be implemented in other ways. The system embodiments described above are merely illustrative, and for example, the flowchart and block diagrams in the figures illustrate the architecture, functionality, and operation of possible implementations of systems, methods and computer program products according to various embodiments of the present invention. In this regard, each block in the flowchart or block diagrams may represent a module, segment, or portion of code, which comprises one or more executable instructions for implementing the specified logical function(s). It should also be noted that, in some alternative implementations, the functions noted in the block may occur out of the order noted in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending upon the functionality involved. It will also be noted that each block of the block diagrams and/or flowchart illustration, and combinations of blocks in the block diagrams and/or flowchart illustration, can be implemented by special purpose hardware-based systems which perform the specified functions or acts, or combinations of special purpose hardware and computer instructions.
In addition, the functional modules in the embodiments of the present invention may be integrated together to form an independent part, or each module may exist alone, or two or more modules may be integrated to form an independent part.
The functions, if implemented in the form of software functional modules and sold or used as a stand-alone product, may be stored in a computer-readable storage medium. Based on such understanding, the technical solution of the present invention may be embodied in the form of a software product, which is stored in a storage medium and includes instructions for causing a computer device (which may be a personal computer, a server, or a network device) to execute all or part of the steps of the method according to the embodiments of the present invention. And the aforementioned storage medium includes: a U-disk, a removable hard disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk or an optical disk, and other various media capable of storing program codes.
In light of the foregoing description of the preferred embodiment of the present invention, many modifications and variations will be apparent to those skilled in the art without departing from the spirit and scope of the invention. The technical scope of the present invention is not limited to the content of the specification, and must be determined according to the scope of the claims.
Claims (10)
1. A method for regulating heat of a heating system heating station based on secondary side flow is characterized by comprising the following steps:
s1, acquiring historical operating data and weather data of each heating power station, and establishing a load prediction model of each heating power station after meteorological pattern clustering, characteristic importance evaluation and model training are carried out on the acquired data through a prediction model;
s2, when the heat load change of each heating power station exceeds a threshold value, calculating the required secondary side water supply flow according to a heat load predicted value;
s3, establishing plate exchange models of heat exchangers of all heating power stations;
s4, establishing adjusting parameters of a secondary side circulating pump and an auxiliary adjusting device of each thermal power station and a corresponding data driving model comprising the water supply flow and temperature at the primary side and the heat exchange quantity between the primary side and the secondary side based on a plate exchange model based on an LSTM model of a double attention mechanism;
s5, adjusting the secondary side circulating pumps and the auxiliary adjusting devices of the thermal power stations according to the data driving model to meet the secondary side water supply flow demand required by the thermal power stations and change the heat of the primary side entering the secondary side;
and S6, analyzing the operation condition after the secondary side flow is adjusted according to the established heat supply system secondary network simulation model, and then correcting the adjusting parameters of the secondary side circulating pump and the auxiliary adjusting device of each heating power station to obtain the optimal adjusting parameters of the secondary side circulating pump and the auxiliary adjusting device.
2. A heating system thermal station heat conditioning method according to claim 1, wherein said step S1 comprises:
respectively acquiring historical operating data and weather data according to a heat meter arranged in a heating station and a weather data interface butted with a heating system, wherein the historical operating data at least comprises secondary side water supply flow, secondary side water supply and return temperature, secondary side water supply and return pressure and heat load, and the weather data at least comprises temperature, humidity, wind speed and illumination;
taking the historical operating data and the weather data as model samples, clustering GMMs by adopting a Gaussian mixture model, and carrying out meteorological feature clustering analysis according to the self properties of the historical weather data to obtain a plurality of meteorological modes;
and selecting the characteristics of the data in the model sample by adopting a random forest algorithm, selecting a characteristic subset with higher characteristic importance, inputting the selected characteristic subset into an optimized LSSVM model, training the data in each meteorological mode, establishing a thermal station load prediction model, and superposing corresponding load prediction results in each meteorological mode to obtain a final thermal station load prediction result.
3. A method for regulating heat of a heating system thermal power station according to claim 2, wherein the implementation of selecting the feature subset with higher feature importance by using a random forest algorithm to perform feature selection on the data in the model sample comprises:
the random forest comprises a plurality of decision trees, the feature importance is calculated according to the contribution rate of each feature in each decision tree, and the feature importance of the feature is obtained by averaging the contribution rates of one feature in all the decision trees; the contribution rate is obtained by calculating a kini coefficient, the feature importance of the jth feature in the node a is calculated according to the change of the kini index, and the contribution rate is expressed as: VIM ja ＝GI a -GI b -GI c ；GI a 、GI b And GI c The node a and the node a are respectively used for generating a Gini coefficient of two new nodes b and c after branching;
assuming a total of n trees in a random forest, the importance of the jth feature over all trees is representedComprises the following steps:
is the sum of the importance of features on n trees; VIM ij Is the sum of the importance of the jth feature in the ith tree;
averaging the sum of the importance of the jth feature, which is the feature importance of the jth feature,
p =1,2,3, \ 8230, n, which is the sum of all the importance of m features on n trees;
sorting all the features from large to small in feature importance, and selecting the features with high feature importance, namely the top n ranked features as a feature subset;
when the GMM model carries out meteorological mode clustering, parameter estimation of mean value and covariance of the initialized GMM is carried out by adopting an expectation maximization algorithm; the expectation maximization algorithm comprises the expectation steps of: the cluster number of the GMM model needs to be set, the pre-estimated values of the mean value and the covariance of the initialized GMM are solved, and the probability that the weather data only belong to the corresponding cluster is calculated; a maximization step: dividing the data points into clusters with higher probability by using a maximum likelihood function, and updating the mean value and covariance of GMM; finally, circularly performing the expectation step and the maximization step until the parameters are converged or the likelihood function is converged to obtain a clustering result of the meteorological model;
the optimized LSSVM model adopts a metaheuristic optimization algorithm AOA to optimize a kernel parameter sigma and a regularization parameter gamma of the LSSVM model, and comprises the following steps: initializing parameters of an AOA optimization algorithm, including population number, maximum iteration number, local development precision and an acceleration function; randomly generating a population, setting initial position parameters (sigma, gamma) of the population, calculating individual fitness values according to the root-mean-square error, and then comparing the fitness values to obtain the current optimal population position; judging whether the initial population enters an exploration phase or a development phase, and updating the position of the initial population; comparing the updated population, taking the population with the lowest fitness as an optimal population position, and judging whether an iteration condition is met; and performing model prediction by taking the optimal value generated by the iteration end as the parameter (sigma, gamma) of the LSSVM.
4. A heating system thermal station heat conditioning method according to claim 1, wherein said step S2 comprises: when the predicted heat load value of each heating power station is compared with the current heat load value, and when the change of the heat load value exceeds a set threshold value, calculating the required secondary side water supply flow according to the predicted heat load value and a set secondary side water supply and return temperature value; otherwise, the current system operation condition is maintained.
5. A method for regulating heat quantity of a heat supply system heat station according to claim 1, wherein in the step S3, establishing a plate exchange model of each heat station heat exchanger comprises: the method comprises the steps of training a plate heat exchange model by adopting a neural network algorithm, and establishing a heat exchange quantity model of the heat exchanger by fitting the heat station heat exchanger with different secondary side water supply flow, different heat exchange quantity and different primary side water supply flow and water supply temperature data, wherein the heat exchange quantity model is used for describing the relationship between the secondary side water supply flow and the heat acquired from the primary side through the heat exchanger under the condition of giving the primary side water supply flow and the water supply temperature.
6. A heating system heat station heat conditioning method according to claim 1, characterized in that said step S4 comprises:
through the different secondary side circulating pump of each heating power station of fitting and the corresponding primary side water supply flow and temperature of auxiliary regulating device regulating parameter, the heat that the secondary side was obtained from the primary side through the heat exchanger and other historical data that influence the relevant parameter of secondary side flow, establish secondary side flow control model for the description is at certain model input vector: the method comprises the following steps that under the conditions of primary side water supply flow and temperature, heat obtained by a secondary side from a primary side through a heat exchanger and other related parameters influencing secondary side flow, a model outputs required secondary side flow adjusting parameters including adjusting parameters of a secondary side circulating pump and an auxiliary adjusting device;
the auxiliary adjusting device comprises a water mixing device arranged on the secondary side of the thermal station, the flow of the water mixing pipe is adjusted by adjusting the opening degree of a valve of the water mixing device, the partial flow of the secondary side backwater is introduced into the water mixing pipe, the secondary side circulating flow is adjusted by adjusting the parameters of a secondary side circulating pump, and the secondary side water supply flow is adjusted by the secondary side circulating pump and the water mixing device together; the other parameters related to the secondary side flow comprise a secondary side water supply and return pressure difference, a secondary side circulating pump attribute, an auxiliary adjusting device attribute and a secondary side water supply and return pressure;
the secondary side flow control model training adopts an LSTM model based on a dual attention mechanism, and comprises the following steps: inputting vector layer, feature attention layer, LSTM network, time sequence attention layer and full connection layer output; the input vector layer obtains an input feature vector x consisting of a secondary side flow regulation historical sequence and a related input feature sequence t And transmitting the input feature vector x to the feature attention layer, and performing feature extraction through dynamic distribution of feature attention weight to obtain weighted and corrected input feature vector x t '; constructing an LSTM network layer structure, extracting hidden time sequence correlation information from the weighted and corrected input feature vector, and obtaining a hidden layer state h of each historical moment t (ii) a Mining the correlation between the information of each moment of the relevant characteristic time sequence and the current moment data through the time sequence attention layer, and calculating the influence weight l of the output information of each historical moment by adopting a time sequence attention mechanism t Finally, the global hidden layer state h of each historical time information is used t Inputting the data to a full connection layer, and outputting a secondary side flow regulation predicted value y of n steps in the future t+n 。
7. A heating system thermal station heat regulation method according to claim 6, wherein the characteristic attention layer inputs a characteristic vector x t As a feature attention mechanismInputting, performing attention weight calculation on m features at the current time, and expressing as follows: e.g. of a cylinder t ＝σ(W e x t +b e )；x t ＝[x 1,t ,x 2,t ,...,x m,t ]；e t ＝[e 1,t ,e 2,t ,...,e m,t ](ii) a σ () is a Sigmoid activation function; w is a group of e Is a trainable weight matrix; b is a mixture of e A bias vector for computing feature attention weights;
normalizing each feature attention weight coefficient to obtain a feature attention weight alpha t ＝[α 1,t ,α 2,t ,...,α m,t ]The attention weight value of the mth feature is expressed as:
feature attention weight alpha obtained by the current time instant t And the input feature vector x t Computing a weighted modified input feature vector x t ', is represented as: x is the number of t ′＝α t ⊙x t ＝[α 1,t x 1,t α 2,t x 2,t … α m,t x m,t ](ii) a An h-hadamard product;
the feature attention layer includes, in addition to a feature attention mechanism, a CNN network for performing local feature extraction on input data before calculating attention weights.
8. A heating system thermal station heat regulation method according to claim 6, characterized in that the temporal attention layer has a hidden layer state h of LSTM network at time t t As an input to the time-series attention mechanism, a time attention weight corresponding to each historical time at the current time is calculated, and is expressed as: l. the t ＝ReLU(W d h t +b d )；h t ＝[h 1,t ,h 2,t ,...,h k,t ](ii) a k is the input sequence time window length; l t ＝[l 1,t ,l 2,t ,...,l k,t ](ii) a ReLU () is an activation function; w d For trainingA weight matrix; b d A bias vector for computing a temporal attention weight;
normalizing each time attention weight coefficient to obtain a time attention weight beta t ＝[β 1,t ,β 2,t ,...,β k,t ]The attention weight value at the kth time is expressed as:
temporal attention weight β obtained by the current time instant t And hidden layer state h t Computing a global hidden layer state h t ', indicated as:
is a matrix product.
9. A method for adjusting heat quantity of a heating power station of a heating system according to claim 1, wherein in step S6, after analyzing the operation condition of the secondary side flow quantity after adjustment according to the established simulation model of the secondary side network of the heating system, the adjusting parameters of the secondary side circulating pump and the auxiliary adjusting device of each heating power station are corrected to obtain the optimal adjusting parameters of the secondary side circulating pump and the auxiliary adjusting device, and the method comprises:
analyzing the system operation hydraulic working condition after flow regulation is carried out according to the regulation parameters of the secondary side circulating pump and the water mixing device according to a pre-established simulation model of the secondary side network of the heat supply system, and correcting the regulation parameters of the secondary side circulating pump and the auxiliary regulation device of each heating power station according to hydraulic power dispatching loss to obtain the optimal regulation parameters of the secondary side circulating pump and the auxiliary regulation device;
the hydraulic power loss scheduling is the ratio of the actual secondary side water supply flow to the predicted secondary side water supply flow, and if the hydraulic power loss scheduling is larger than 1, the actual secondary side water supply flow of the heating power station is larger than the predicted flow; if the hydraulic power loss schedule is less than 1, the actual secondary side water supply flow of the heating power station is less than the predicted flow;
adjusting and correcting different floating values of adjusting parameters of each heating power station secondary side circulating pump and each auxiliary adjusting device according to hydraulic power failure scheduling to obtain multiple adjusting schemes, taking the adjusting parameters in different adjusting schemes as control quantities, outputting heating power station secondary side water supply flow corresponding to different adjusting schemes after the adjusting parameters are operated through a constructed secondary network simulation model, taking the variance of the heating power station secondary side water supply flow as a hydraulic power balance standard, and calculating by adopting a global optimization algorithm to obtain an optimal adjusting scheme.
10. A heating system heating power station heat governing system based on secondary side flow, its characterized in that, heating system heating power station heat governing system includes:
the heat load prediction module is used for acquiring historical operating data and weather data of each heat station, and establishing a load prediction model of each heat station after meteorological pattern clustering, characteristic importance evaluation and model training are carried out on the acquired data through the prediction model;
the secondary side flow demand module is used for calculating the required secondary side water supply flow according to the heat load predicted value when the heat load change of each heating power station exceeds a threshold value;
the plate exchange model establishing module is used for establishing a plate exchange model of each heat station heat exchanger;
the secondary side flow control module is used for establishing adjusting parameters of a secondary side circulating pump and an auxiliary adjusting device of each thermal power station and a corresponding data driving model comprising primary side water supply flow and temperature and heat exchange quantity between the primary side and the secondary side based on the plate exchange model based on an LSTM model of a double attention mechanism;
the execution module is used for adjusting the secondary side circulating pump and the auxiliary adjusting device of each heating power station according to the data driving model, meeting the secondary side water supply flow demand value required by each heating power station and changing the heat of the primary side entering the secondary side;
and the adjusting parameter correcting module is used for correcting the adjusting parameters of the secondary side circulating pump and the auxiliary adjusting device of each heating power station after analyzing the operation condition of the secondary side flow adjusted according to the established heat supply system secondary network simulation model to obtain the optimal adjusting parameters of the secondary side circulating pump and the auxiliary adjusting device.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211298175.6A CN115751441A (en) | 2022-10-22 | 2022-10-22 | Heat supply system heating station heat regulation method and system based on secondary side flow |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211298175.6A CN115751441A (en) | 2022-10-22 | 2022-10-22 | Heat supply system heating station heat regulation method and system based on secondary side flow |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115751441A true CN115751441A (en) | 2023-03-07 |
Family
ID=85352709
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211298175.6A Pending CN115751441A (en) | 2022-10-22 | 2022-10-22 | Heat supply system heating station heat regulation method and system based on secondary side flow |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115751441A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116839091A (en) * | 2023-05-15 | 2023-10-03 | 山东和同信息科技股份有限公司 | Heat exchange station automatic control parameter setting method based on deep learning |
| CN117870008A (en) * | 2023-12-14 | 2024-04-12 | 华能济南黄台发电有限公司 | Intelligent big data driven heat supply energy-saving optimization management method and device |
| CN117968143A (en) * | 2024-04-01 | 2024-05-03 | 陕西德联新能源有限公司 | Energy-saving optimization method and system for heating system |
-
2022
- 2022-10-22 CN CN202211298175.6A patent/CN115751441A/en active Pending
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116839091A (en) * | 2023-05-15 | 2023-10-03 | 山东和同信息科技股份有限公司 | Heat exchange station automatic control parameter setting method based on deep learning |
| CN116839091B (en) * | 2023-05-15 | 2024-01-19 | 山东和同信息科技股份有限公司 | Heat exchange station automatic control parameter setting method based on deep learning |
| CN117870008A (en) * | 2023-12-14 | 2024-04-12 | 华能济南黄台发电有限公司 | Intelligent big data driven heat supply energy-saving optimization management method and device |
| CN117968143A (en) * | 2024-04-01 | 2024-05-03 | 陕西德联新能源有限公司 | Energy-saving optimization method and system for heating system |
| CN117968143B (en) * | 2024-04-01 | 2024-06-04 | 陕西德联新能源有限公司 | Energy-saving optimization method and system for heating system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111561734B (en) | Heating household valve adjusting method, system and equipment based on artificial intelligence | |
| CN107944648B (en) | Large ship speed and oil consumption rate prediction method | |
| CN111148118A (en) | Flow prediction and carrier turn-off method and system based on time sequence | |
| CN105701502A (en) | Image automatic marking method based on Monte Carlo data balance | |
| CN115751441A (en) | Heat supply system heating station heat regulation method and system based on secondary side flow | |
| Basurto et al. | A hybrid intelligent system to forecast solar energy production | |
| US20220282884A1 (en) | Machine learning device, demand control system and air-conditioner control system | |
| CN111461466B (en) | Heating valve adjusting method, system and equipment based on LSTM time sequence | |
| CN114022311B (en) | Comprehensive energy system data compensation method for generating countermeasure network based on time sequence condition | |
| CN114811713B (en) | Two-level network inter-user balanced heat supply regulation and control method based on mixed deep learning | |
| CN112232577A (en) | Power load probability prediction system and method for multi-core intelligent meter | |
| CN112766603A (en) | Traffic flow prediction method, system, computer device and storage medium | |
| CN116538127B (en) | Axial flow fan and control system thereof | |
| CN111160461A (en) | Fuzzy clustering-based weighted online extreme learning machine big data classification method | |
| CN117313795A (en) | Intelligent building energy consumption prediction method based on improved DBO-LSTM | |
| CN113344243A (en) | Wind speed prediction method and system for optimizing ELM based on improved Harris eagle algorithm | |
| CN114330119B (en) | Deep learning-based extraction and storage unit adjusting system identification method | |
| CN115409369A (en) | Comprehensive energy system reliability evaluation method based on mechanism and data hybrid driving | |
| CN112183721B (en) | Construction method of combined hydrological prediction model based on self-adaptive differential evolution | |
| CN113033898A (en) | Electrical load prediction method and system based on K-means clustering and BI-LSTM neural network | |
| CN110163437A (en) | Day-ahead photovoltaic power generation power prediction method based on DPK-means | |
| CN116435987A (en) | Ultra-short-term wind power prediction method for large-scale wind power clusters | |
| CN116227704A (en) | Edge calculation and cloud calculation-based heat supply energy consumption diagnosis and energy-saving regulation and control method | |
| CN113420492A (en) | Modeling method for frequency response model of wind-solar-fire coupling system based on GAN and GRU neural network | |
| KR102497543B1 (en) | Military demand prediction model and practical system using machine learning |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |